RESEARCHARTICLE

Predicting 10-Year Risk of FatalCardiovascular Disease in Germany: AnUpdate Based on the SCORE-DeutschlandRisk ChartsViktoriaRücker1, Ulrich Keil2, Anthony P Fitzgerald3, UweMalzahn1,4, ChristofPrugger5,Georg Ertl6,7, Peter U Heuschmann1,4,7, HanneloreNeuhauser8,9*

1 Institute of Clinical Epidemiology and Biometry, University of Würzburg, Würzburg, Germany, 2 Institute ofEpidemiology and Social Medicine, University of Münster, Münster, Germany, 3 Department of Epidemiologyand Public Health, Department of Statistics, University College Cork, Cork, Ireland,4 Clinical Trial Center,University HospitalWürzburg, Würzburg, Germany, 5 Institute of Public Health, Charité‒Universitätsmedizin Berlin, Berlin,Germany, 6 Department of Medicine I, University HospitalWürzburg,Würzburg, Germany, 7 Comprehensive HeartFailure CenterWürzburg, UniversityWürzburg, Würzburg,Germany, 8 Department of Epidemiology and HealthMonitoring, RobertKoch InstituteBerlin, Berlin,Germany, 9 DZHK (GermanCenter for Cardiovascular Research), partnersite Berlin, Berlin,Germany

* neuhauserh@rki.de

AbstractEstimation of absolute risk of cardiovascular disease (CVD), preferably with population-spe-

cific risk charts, has become a cornerstoneof CVD primaryprevention. Regular recalibra-

tion of risk chartsmay be necessary due to decreasing CVD rates and CVD risk factor

levels. The SCORE risk charts for fatal CVD risk assessment were first calibrated for Ger-

many with 1998 risk factor level data and 1999 mortalitystatistics. We present an update of

these risk charts based on the SCOREmethodology including estimates of relative risks

fromSCORE, risk factor levels from the GermanHealth Interview and Examination Survey

for Adults 2008–11 (DEGS1) and official mortality statistics from 2012. Competing risks

methods were applied and estimates were independently validated. Updated risk charts

were calculated based on cholesterol, smoking, systolic blood pressure risk factor levels,

sex and 5-year age-groups. The absolute 10-year risk estimates of fatal CVD were lower

according to the updated risk charts compared to the first calibration for Germany. In a

nationwide sample of 3062 adults aged 40–65 years free of major CVD fromDEGS1, the

mean 10-year risk of fatal CVD estimated by the updated chartswas lower by 29% and the

estimated proportionof high risk people (10-year risk > = 5%) by 50% compared to the older

risk charts. This recalibration shows a need for regular updates of risk charts according to

changes in mortalityand risk factor levels in order to sustain the identification of people with

a high CVD risk.

PLOSONE | DOI:10.1371/journal.pone.0162188 September 9, 2016 1 / 11

a11111

OPENACCESS

Citation:Rücker V, Keil U, Fitzgerald AP, Malzahn U,Prugger C, Ertl G, et al. (2016) Predicting 10-YearRisk of Fatal Cardiovascular Disease in Germany: AnUpdate Based on the SCORE-Deutschland RiskCharts. PLoS ONE 11(9): e0162188. doi:10.1371/journal.pone.0162188

Editor: Chiara Lazzeri, AziendaOspedalieroUniversitaria Careggi, ITALY

Received:April 29, 2016

Accepted:August 18, 2016

Published:September 9, 2016

Copyright:© 2016 Rücker et al. This is an openaccess article distributed under the terms of theCreative Commons Attribution License, which permitsunrestricteduse, distribution, and reproduction in anymedium, provided the original author and source arecredited.

Data Availability Statement: The authors confirmthat some access restrictions apply to the dataunderlying the findings. The informed consent givenby DEGS1 study participantsand by studyparticipantsof the cohorts forming the SCORE cohortdoes not cover posting individual data in publicdatabases.However, DEGS 1 data are available uponrequest by means of project agreements.Datarequests can be sent to the ElectronicData Center ofthe Department of Epidemiologyand Health Reportingof the Robert Koch Institute at EDS@rki.de.

IntroductionCurrent guidelines on prevention of cardiovascular disease (CVD) such as the EuropeanGuidelines on cardiovascular disease prevention in clinical practice [1], recommend to guidedecisions to take preventive actions as well as the level of preventive actions (e.g. lifestyle rec-ommendations, drug treatment) by an estimation of absolute risk of CVD based on multiplecardiovascular risk factors [2, 3]. Since CVD prevention is not only an individual concern butalso a worldwide public health concern, there is a need for a common, albeit locally specificrisk assessment that can take account of regional disparities in CVD risk as well as of timetrends in CVD mortality and CVD risk factors.

As a result, the Third Joint Task Force of European and other Societies on CardiovascularDisease Prevention in Clinical Practice proposed in 2003 new CVD risk charts for Europe andsuggested the SCORE (Systematic COronary Risk Evaluation) project to estimate ten-year riskof fatal cardiovascular disease [4, 5]. The SCORE charts were intended to be calibrated locallywhich is an advantage compared to non-European based risk equations which considerablyoverestimate absolute risk in European populations [6–8]. In the following years, risk chartsfor high-risk and low-risk European countries have been issued and in addition national orregional calibrations of the SCORE risk charts were performed in order to reflect specific mor-tality and risk factor levels [4, 9–12]. These risk charts should be used for risk estimation in pri-mary prevention only since patients with clinically manifest CVD are already considered ashigh-risk individuals [1, 4]. In addition, risk can be higher than indicated in the risk charts inpatients with specific additional comorbidities (e.g. diabetes) as detailed in CVD preventionguidelines. [1]. Accordingly, in 2005, a recalibrated SCORE-Deutschland risk chart was pub-lished based on data for risk factor levels from the German National Health Interview andExamination Survey 1998 (GNHIES98) and the official national mortality statistics of 1999from the Federal Statistical Office (Destatis) [13]. Since then, decreasing trends in major CVDrisk factors such as decreasing blood pressure and cholesterol levels [14, 15] and to a lesserextent smoking prevalence [16] have been observed in Germany, as well as a marked decreasein CVD mortality rates [17]. These trends suggested the need for an update of the risk chartsfor Germany in order to avoid overestimation of risk and false positive categorization of peoplefrom the general population into the high risk group.

Therefore, we updated the calibration of the risk charts based on the SCORE methodologyfor Germany using risk factor profiles from the most recent German National Health Examina-tion survey (DEGS1 study 2008–2011) as well as national CVD mortality rates in 2012. Inaddition, risk estimates based on the risk charts published in 2005 and the updated ones werecompared in a representative sample of the general population in Germany.

Methods

Score projectThe first SCORE risk charts for Europe were published in 2003 [4]. The SCORE Project pooleddata from 12 European prospective cohort studies from the years 1967–1991 including theGerman MONICA-Augsburg cohort [18]. The SCORE dataset contains information from205,178 persons (43% women and 57% men) representing 2.7 million person years of follow-up and 7,934 documented cardiovascular deaths. Based on this dataset, risk charts for cardio-vascular mortality were developed in 2003 predicting the absolute risk of dying from CVD inthe next ten years. The risk was modelled as a function of age, sex, systolic blood pressure(SBP), cholesterol (total (TC) or total cholesterol/HDL-ratio (TC/HDL-C)) and smoking.Two SCORE risk charts were developed, one for European populations at a high CVD risk

Risk Chartsof Cardiovascular Disease in Germany

PLOSONE | DOI:10.1371/journal.pone.0162188 September 9, 2016 2 / 11

Funding: The work of VR has been supported by theGerman Centre of Cardiovascular Research, DZHK,grant MDC2014Z3 (http://dzhk.de/en/). The DZHK isfunded by the Federal Ministry of Education andResearch (BMBF). The funders had no role in studydesign, data collection and analysis, decision topublish, or preparationof the manuscript.

Competing Interests: I have read the journal's policyand the authors of this manuscript have the followingcompeting interests: HN, VR, UK, CP declare noconflict of interest. PUH reports research grants fromthe GermanMinistry of Education and Research,EuropeanUnion, Charité, Berlin Chamber ofPhysicians, German Parkinson Society, UniversityHospitalWürzburg, Robert-Koch-Institute, Charité–Universitätsmedizin Berlin (within MonDAFIS;MonDAFIS is supported by an unrestricted researchgrant to the Charité from Bayer), University Göttingen(within FIND-AFrandomized; FIND-AFrandomized issupported by an unrestricted research grant to theUniversity Göttingen from Boehringer-Ingelheim), andUniversity Hospital Heidelberg (within RASUNOA-prime; RASUNOA-prime is supported by anunrestricted research grant to the University HospitalHeidelberg from Bayer, BMS, Boehringer-Ingelheim,Daiichi Sankyo), outside submittedwork. GE reportsresearch grants from the German Ministry ofEducation and Research (BMBF), Lundbeck non-financial support, Bayer grants, Novartis grants andpersonal fees. This does not alter our adherence toPLOS ONE policies on sharing data and materials.

(SCORE-high) and one for European populations at a low CVD risk (SCORE-low). For a moreprecise prediction of risk estimates, calibration of the risk charts according to risk factor andmortality levels of respective countries was proposed [4].

First calibration for GermanyThe first calibration for Germany was published in 2005 [13]. The SCORE-Deutschland riskcharts were based on three terms: a German specific baseline absolute risk of death from CVD,relative risks based on the estimates from the SCORE Project and estimates of average risk fac-tor levels of SBP, cholesterol and smoking. The calculations of the SCORE-Deutschland riskcharts were based on a Cox proportional hazards model with age as a time variable withdelayed entry. Both TC and TC/HDL-C based SCORE-Deutschland risk charts were madeavailable. Previous analyses of the SCORE data had shown similar performance of the TC andthe TC/HDL-C based charts and risk estimation with any of the two versions had been recom-mended in the first German SCORE calibration [13]. The calculations of the absolute risk forthe SCORE-Deutschland risk charts published in 2005 were calibrated using the official mor-tality rates of the year 1999 and risk factor data from the GNHIES98 [19, 20]; the relative riskswere derived from the SCORE dataset [4].

Updated calibration for GermanyFor the updated calibration of 10-year risk of CVD based on the SCORE methodology, recentnational estimates of age- and sex-specific mean levels of the risk factors SBP, smoking andcholesterol (TC or TC/HDL-C) from the DEGS1 2008–2011 study were used. The investiga-tions of the DEGS1 study were carried out in accordance with the Declaration of Helsinki,including written informed consent of all participants. All study methods were approved bythe ethics committee of Charité–Universitätsmedizin Berlin and the federal data protectionand deral data protection and freedom of information authority. The examination sample ofDEGS1 included 7.115 participants aged 18–79 years from a Germany-wide two-stage clus-tered sample from local population registries with standardized measurements and interviews[21]. TC and high density lipoprotein cholesterol (HDL-C) were measured using an enzymaticassay (Architect ci8200, Abbott, Germany). Non-HDL-C was calculated by subtracting HDL-Cfrom TC. Three blood pressure (BP) measurements were taken at the right arm at 3-minuteintervals after a non-strenuous part of the examination and an additional 5-minute rest withan automated oscillometric device (Datascope Accutorr Plus, Mahwah, NJ). Participants sat ona height-adjustable chair, back supported, arm lying on a table at heart level, feet on the floorand legs uncrossed. Three different cuff sizes were used. The second and third measurementswere averaged [14]. Smoking was defined as current self-reported smoking. In order to accountfor the unequal sampling probabilities related to the sampling design and nonresponse, statisti-cal analyses were weighted using a weighting factor [21] and the complex survey sampling wastaken into account using complex samples procedures in SPSS version 20.0 (SPSS Inc., Chi-cago, Illinois, USA).

The updated calibration also used recent mortality data from Germany. Sex-stratified and5-year age aggregated number of deaths from CVD and non-CVD (CVD: ICD-10 I10-I15,I20-I25, I44-I51 and I61-I73) and the number of inhabitants from the year 2012 were collectedfrom official mortality statistics. We used competing risk model that utilizes the cumulativeincidence function to calculate the average absolute risk of dying from CVD. The competingrisks model allows for the fact that there are competing causes of death that influence theactual likelihood of dying from CVD. Non-CVD death was considered as being the competingevent and death from CVD the event of interest. It was assumed that the relative risks of the

Risk Chartsof Cardiovascular Disease in Germany

PLOSONE | DOI:10.1371/journal.pone.0162188 September 9, 2016 3 / 11

respective risk factors didn’t change over the years. Therefore, relative risks for CVD mortalitywere derived from an analysis of the entire SCORE database [4]. The statistical method for theupdated calibration is based on the methodology introduced previously [13] but was slightlymodified as follows: The calibration from 2005 interpolated the 5-year mortality rates to getthe average absolute risk of dying in the next ten years. In the new approach the 5-year mortal-ity rates were held constant. No interpolation was used, because a linear interpolation disagreeswith the Markov assumption. Therefore, the estimates of age- and sex-specific mean levels ofthe risk factors were evaluated with regression models using age in constant 5-year age groupsand not as a continuous variable. Thus, separate risk estimations for the 40–45 and for the 45–49 year olds were provided in contrast to the 2005 calibration. As a sensitivity analysis, an inde-pendent recalculation of the risk charts using the original approach was performed by one ofthe authors who was blinded to the current recalibration analysis (APF). Statistical analyseswere performed in R 3.1.1. A detailed description of the recalibration methods is provided inthe Supporting information (S1 File).

Changes in 10-year risk of fatal CVD estimationThe change in estimated risk between the first SCORE-Deutschland risk charts published in2005 and the updated risk charts presented in this paper has been evaluated by applying thetwo risk charts to a nationwide sample of community-dwelling adults aged 40–65 years fromthe DEGS1 study in 2008–2011 [21]. Out of 3,470 participants of the DEGS1 study aged 40 to65 years, 131 (3.8%) were excluded due to missing data on major CVD (self-reported previousdiagnoses of coronary heart disease (CHD), stroke and heart failure) or missing data on vari-ables included in SCORE. Of the remaining participants, 8.1% (men 10.9%, women 5.1%) wereexcluded due to a history of major CVD, leaving a sample of 3,061 participants (unweightedsample size) for estimation of cardiovascular risk (49.5% women, 50.5% men, mean age ± SD50.9 ± 7.0 years). 10-year risks of fatal CVD were calculated using both the original and theupdated calibration methods. For higher precision, risk was directly calculated from the respec-tive formulas not from risk chart readings. High risk was defined according to current guide-lines as estimated 10-year risk of fatal CVD of 5% or more. Statistical analyses of DEGS1 datawere weighted as described above.

ResultsThe updated risk charts for estimation of the absolute 10-year- risk of fatal CVD in Germanyare presented in Figs 1 and 2. Two versions are shown, one using TC (Fig 1) and one using TC/HDL-C (Fig 2). Both versions are stratified by sex and include age, smoking and SBP as addi-tional risk factors. To estimate the absolute 10-year risk of CVD death for an individual, his orher age needs to be rounded to the nearest age shown on the chart and the smoker or non-smoker cell nearest to the person’s SBP and TC or TC/HDL-C ratio will provide the estimatedrisk.

The updated charts consist of 480 cells. For comparison the cells of the age group 40–49from 2005 were compared to the cells of age group 40–44 and to the ones of the age group 45–49 from the updated risk chart. Compared to the 2005 predictions in the tables with TC, 34.2%of the absolute cell predictions were higher and 5.6% were lower. For example, the highestabsolute risk displayed in a cell of the updated risk charts compared to the one published in2005 decreased in the TC chart from 37% to 34% for men (absolute risk difference of 3) andfrom 19% to 15% for women (absolute risk difference of 4) in the highest category for all riskfactors. The differences between the old and the updated risk chart for TC in correspondingcells for the age categories 50–65 had a mean of 0.69% (SD +/- 0.72). For the tables with TC/

Risk Chartsof Cardiovascular Disease in Germany

PLOSONE | DOI:10.1371/journal.pone.0162188 September 9, 2016 4 / 11

HDL-C, 47.1% of the absolute risks estimated for Germany decreased in comparison with the2005 risk chart [13] and the highest absolute risks displayed in a cell had a risk difference of 4percentage points in the charts for men and 5 for women.

Since a slight adaptation in the calibration method was introduced compared to the first cal-ibration of 2005 (use of constant vs. interpolated 5-year mortality rates), a sensitivity analysiswas conducted and the updated calibration was performed using both methods. This sensitivityanalysis showed an excellent agreement (intraclass correlation coefficient ICC = 0.99) betweenthe method used in the 2005 calibration and the present approach. Agreement was calculatedfor the age range 50–65 years (320 cells) since for younger ages the cells do not directly

Fig 1. Updated (2015) risk charts for estimationof absolute 10-year risk (%) of fatal CVD in Germany based on the SCORE-Deutschland riskcharts (tableswith total cholesterol).

doi:10.1371/journal.pone.0162188.g001

Risk Chartsof Cardiovascular Disease in Germany

PLOSONE | DOI:10.1371/journal.pone.0162188 September 9, 2016 5 / 11

correspond (the first calibration has the age category 40–49, the updated calibration has thetwo age categories 40–44 and 45–49). The differences between the two methods in correspond-ing cells had a mean of 0.24% (SD +/- 0.27). The risk predictions using the method from 2005were always slightly higher than the ones using the method from 2015. The method used in2005 classified 3 additional cells (out of 320) as high risk compared to the method used in2015.

When applied to a representative sample of the population living in Germany aged 40 to 65years without a previous history of major CVD from the national health survey DEGS1 2008–2011, the updated risk charts using TC classified significantly fewer men and women as high-

Fig 2. Updated (2015) risk charts for estimation of absolute10-year risk (%) of fatal CVD in Germany based on the SCORE-Deutschland riskcharts (tableswith total cholesterol/HDL cholesterol ratio).

doi:10.1371/journal.pone.0162188.g002

Risk Chartsof Cardiovascular Disease in Germany

PLOSONE | DOI:10.1371/journal.pone.0162188 September 9, 2016 6 / 11

risk (10-year risk> = 5%) than the German 2005 charts (German risk charts published 2005:5.2% [95% CI 4.3–6.1]; updated risk charts: 2.6% [95% CI 2.0–3.3]). Average risk was also sig-nificantly lower in all age groups among men and women according to the updated chartscompared to the charts from 2005 (German risk charts published in 2005: 1.47% [95% CI1.40–1.54]; updated risk charts 1.04% [95% CI 0.99–1.09]) (Fig 3).

DiscussionThe SCORE risk charts for estimating the absolute 10-year risk of fatal CVD for the Germanpopulation from 2005 [4, 13] have been updated successfully based on the SCORE methodol-ogy [4]. More than a decade after the first calibration of the SCORE charts for Germany [13]these updated charts reflect the decreasing trend in CVD mortality in Germany as well asrecent trends towards lower risk factor levels. As expected, the predicted absolute risk generallydecreased. The use of the calibrated charts from 2005 for Germany showed a substantiallyhigher population risk in a representative sample of the general population living in Germanyaged 40 to 65 years compared to the updated estimates.

Germany has seen a marked decline in cardiovascular mortality in the period between thetwo risk chart calibrations. The calibration published in 2005 and the updated calibrations arebased on mortality data from 1998 and 2012 respectively, a period in which stroke mortalitydeclined by 51%, ischemic heart disease mortality by 45% and overall CVD mortality by 36%(www.gbe-bund.de; mortality rates per 100.000 inhabitants standardized to the population inGermany 2011). In addition, there have been substantial changes in risk factor levels, as docu-mented by the two national health surveys conducted in 1998 and 2008–2011. For example,blood pressure declined considerably (both systolic and diastolic blood pressure by 5 mmHg)[14]. Moreover, cholesterol levels [15] and the proportion of smokers in the German popula-tion decreased slightly [16]. However, the fall in CVD mortality might only be partly explainedby the changes in risk factors as the chart predicts individual risk at specific risk factors levelsand, thus, we wouldn’t expect major changes in the risk chart if the observed decrease in mor-tality would be entirely explained by risk factor changes.

Numerous risk charts exist for estimating CVD risk. [22] In particular newer risk chartsacknowledge the need to issue risk charts for individual countries and recalibrate them regu-larly [19]. The first recalibration of the risk charts for Germany based on the SCORE projectresulted in risk estimates that were between those of SCORE for high-risk and those of SCOREfor low-risk countries [13]. The updated risk charts presented in this paper make furtheradjustments toward lower risk while still lying between SCORE-high and SCORE-low [4].More than 40% of the 480 cells in the risk charts for Germany are higher and around 1% arelower in comparison to the original SCORE-low risk chart cells [4]. In the updated risk chartsthe age group 40–49 is divided into two age groups, 40–44 and 45–49 years. This allows for amore precise prediction of the absolute risk. Predicted risk is generally lower according to theupdated risk charts compared to the first SCORE-Deutschland risk charts. However, due torounding of the predicted risk presented in the chart cells rounded predicted risk is numericallythe same in more than half of the cells.

Other risk charts are also used in Germany such as the PROCAM score [23], which is basedon an occupational cohort from the northwest of Germany, as well as charts like ARRIBA [24]and CARRISMA [25], which are based on adaptations of Framingham risk scores. Thesecharts, however, do not take into account the recent declines in CVD and CVD risk factors inGermany. Results of a German survey on the use of routine risk-scores among physiciansshowed that more than 50% of the participating physicians believe that risk-scores for CVDare useful for the successful prevention and treatment of CVD [26]. Therefore, it is important

Risk Chartsof Cardiovascular Disease in Germany

PLOSONE | DOI:10.1371/journal.pone.0162188 September 9, 2016 7 / 11

Fig 3. Comparison of absolute10-year-risk of fatal CVD by age group in adults aged 40 to 65 years in Germany withoutmajor CVD according to the first vs. the updated risk charts.

doi:10.1371/journal.pone.0162188.g003

Risk Chartsof Cardiovascular Disease in Germany

PLOSONE | DOI:10.1371/journal.pone.0162188 September 9, 2016 8 / 11

that a regular update of the risk-scores based on the current risk factor and mortality distribu-tion of the general population is available.

Our study has certain strengths and limitations. Our recalibration has been confirmed byindependent recalculation. The strengths of risk charts based on the SCORE methodologyinclude the large database of European cohorts and the simple and inexpensive use due torestriction to only a few key risk factors. The original SCORE chart combined separate predic-tions of CHD and stroke mortality.6 The method presented here is based directly on nationalCVD mortality which allows for a more straightforward recalibration. This is of great impor-tance since CVD mortality as well as risk factor levels have been changing considerably in thelast decades in many countries [17, 27]. However, two aspects of the SCORE-recalibration maylead to an overestimation of the absolute CVD risk in the general population without preexist-ing CVD. Ideally, calibration would not require national mortality statistics (which includepeople with previous CVD who have a higher 10-year-mortality risk than those without a pre-vious CVD), but recent and regional data on mortality of a population free of CVD at baseline.This could be approximated from national mortality statistics with recent and regional CVDincidence data, which, however, are currently lacking. The second aspect is that the competingrisk principle is only applied to the country-specific baseline risk but not in the Cox regressionto estimate the hazard ratios. This could be solved by using a Fine and Gray Model with com-peting risks [28] in the 12 original cohorts of the SCORE Project, but is computationally chal-lenging in the presence of left censoring and a model with age as time variable. Finally, it is ageneral limitation of the SCORE charts that they are limited to fatal CVD prediction and to theage range 40 to 65 years. However, the original SCORE data support robust risk estimationonly for fatal CVD and only in this age range. Risk prediction for higher age groups would onlybe possible by adding cohort data in a higher age range as shown in a recent Norwegian exam-ple [29].

ConclusionThe risk charts presented in this paper update the first SCORE-Deutschland CVD risk charts,based on the SCORE methodology taking into account more than a decade of changes inmortality and risk factor levels in Germany. The absolute ten-year risk of CVD in Germanydecreased substantially compared to the risk charts published in 2005. The updated chartsallow more exact risk predictions and can help to improve CVD prevention on the populationlevel by a more precise identification of patients with a high risk of CVD.

Supporting InformationS1 File. Recalibrationmethods.(DOCX)

AcknowledgmentsSupported by the DZHK (German Centre for Cardiovascular Research) and by the BMBF(German Ministry of Education and Research). We thank the German Cardiac Society fortheir general support and Dr. Beyersmann, Ulm, for statistical comments.

Author Contributions

Conceptualization:VR UK APF UM GE PUH HN.

Formal analysis:VR APF.

Risk Chartsof Cardiovascular Disease in Germany

PLOSONE | DOI:10.1371/journal.pone.0162188 September 9, 2016 9 / 11

Funding acquisition: PUH HN.

Methodology:VR APF UM.

Supervision:PUH APF HN.

Visualization: VR HN.

Writing – original draft:VR APF PUH HN.

Writing – review& editing: VR UK APF UM CP GE PUH HN.

References1. Perk J, De Backer G, Gohlke H, Graham I, Reiner Z, VerschurenM, et al. EuropeanGuidelines on car-

diovascular disease prevention in clinical practice (version 2012). The Fifth Joint Task Force of theEuropeanSociety of Cardiology and Other Societies on Cardiovascular Disease Prevention in ClinicalPractice (constitutedby representatives of nine societies and by invited experts). Eur Heart J. 2012; 33(13):1635–701. doi: 10.1093/eurheartj/ehs092 PMID: 22555213.

2. Jackson R. Guidelines on preventing cardiovascular disease in clinical practice. Bmj. 2000; 320(7236):659–61. PMID: 10710556.

3. Jackson R, Lawes CM, Bennett DA, Milne RJ, Rodgers A. Treatment with drugs to lower blood pres-sure and blood cholesterol based on an individual's absolute cardiovascular risk. Lancet. 2005; 365(9457):434–41. doi: 10.1016/S0140-6736(05)17833-7PMID: 15680460.

4. Conroy RM, Pyorala K, Fitzgerald AP, Sans S, Menotti A, De Backer G, et al. Estimation of ten-year riskof fatal cardiovascular disease in Europe: the SCORE project. Eur Heart J. 2003; 24(11):987–1003.PMID: 12788299.

5. De Backer G, Ambrosioni E, Borch-Johnsen K, BrotonsC, Cifkova R, Dallongeville J, et al. Europeanguidelines on cardiovascular disease and prevention in clinical practice. Atherosclerosis. 2003; 171(1):145–55.PMID: 14686332.

6. Hense HW, Schulte H, Lowel H, AssmannG, Keil U. Framingham risk function overestimates risk ofcoronaryheart disease in men and women fromGermany—results from theMONICAAugsburg andthe PROCAM cohorts. Eur Heart J. 2003; 24(10):937–45. PMID: 12714025.

7. ThomsenTF, McGeeD, DavidsenM, JorgensenT. A cross-validation of risk-scores for coronaryheartdiseasemortalitybased on data from the GlostrupPopulation Studies and FraminghamHeartStudy.Int J Epidemiol. 2002; 31(4):817–22. PMID: 12177028.

8. Menotti A, Puddu PE, Lanti M. Comparison of the Framingham risk function-based coronarychartwithrisk function from an Italian population study. Eur Heart J. 2000; 21(5):365–70. PMID: 10666350.

9. Panagiotakos DB, Fitzgerald AP, Pitsavos C, Pipilis A, Graham I, Stefanadis C. Statisticalmodelling of10-year fatal cardiovascular disease risk in Greece: the HellenicSCORE (a calibration of the ESCSCORE project). Hellenic journal of cardiology: HJC = Hellenike kardiologike epitheorese. 2007; 48(2):55–63.Epub 2007/05/11.PMID: 17489342.

10. Marques-Vidal P, Rodondi N, BochudM, Pecoud A, Hayoz D, Paccaud F, et al. Predictive accuracyand usefulness of calibration of the ESC SCORE in Switzerland. Eur J Cardiovasc Prev Rehabil. 2008;15(4):402–8. doi: 10.1097/HJR.0b013e3282fb040f PMID: 18677163.

11. van Dis I, KromhoutD, Geleijnse JM, Boer JM, VerschurenWM. Evaluation of cardiovascular risk pre-dicted by different SCORE equations: the Netherlands as an example. Eur J Cardiovasc Prev Rehabil.2010; 17(2):244–9. doi: 10.1097/HJR.0b013e328337cca2PMID: 20195155.

12. Jdanov DA, Deev AD, Jasilionis D, Shalnova SA, Shkolnikova MA, Shkolnikov VM. Recalibration of theSCORE risk chart for the Russian population. Eur J Epidemiol. 2014; 29(9):621–8. doi: 10.1007/s10654-014-9947-7PMID: 25179794.

13. Keil U, Fitzgerald AP, Gohlke H,Wellmann J, Hense HW. Risk stratificationof cardiovascular diseasesin primaryprevention—The new SCORE-Deutschland Risk Charts.Deutsches Ärzteblatt. 2005; 102(25):1808–12.

14. NeuhauserHK, Adler C, RosarioAS, DiederichsC, EllertU. Hypertensionprevalence, awareness,treatment and control in Germany1998 and 2008–11. J HumHypertens. 2015; 29(4):247–53. doi: 10.1038/jhh.2014.82 PMID: 25273858.

15. Truthmann J, Schienkiewitz A., MensinkG.B.M., BuschM., Knopf H., Du Y., et al. Time trends of totalcholesterol levels 1997–99 to 2008–11 in adults in Germany. Proceedings of the 10thmeeting of theGermanEpidemiological Society. 2015.

Risk Chartsof Cardiovascular Disease in Germany

PLOSONE | DOI:10.1371/journal.pone.0162188 September 9, 2016 10 / 11

16. LampertT, von der Lippe E, Muters S. [Prevalence of smoking in the adult population of Germany:results of the GermanHealth Interviewand Examination Survey for Adults (DEGS1)]. Bundesgesund-heitsblatt GesundheitsforschungGesundheitsschutz. 2013; 56(5–6):802–8. doi: 10.1007/s00103-013-1698-1 PMID: 23703501.

17. Townsend N, Nichols M, Scarborough P, Rayner M. Cardiovascular disease in Europe—epidemiologi-cal update 2015. Eur Heart J. 2015; 36(40):2696–705. doi: 10.1093/eurheartj/ehv428 PMID:26306399.

18. Keil U, Liese AD, Hense HW, Filipiak B, DöringA, Stieber J, et al. Classical risk factors and their impacton incident non-fatal and fatal myocardial infarction and all-cause mortality in southernGermany. EurHeart J 1998; 19(8):1197–207. doi: 10.1053/euhj.1998.1089PMID: 9740341

19. ThammM. [Blood pressure in Germany—current status and trends]. Das Gesundheitswesen. 1999; 61Spec No:S90–3. PMID: 10726403.

20. Junge B, Nagel M. Smoking in Germany. Das Gesundheitswesen. 1999; 61:S121–S5.PMID:10726409

21. Scheidt-Nave C, Kamtsiuris P, Gosswald A, Holling H, LangeM, BuschMA, et al. Germanhealth inter-view and examination survey for adults (DEGS)—design, objectives and implementation of the firstdata collectionwave. BMCPublic Health. 2012; 12:730. doi: 10.1186/1471-2458-12-730 PMID:22938722; PubMed Central PMCID: PMC3490742.

22. Cooney MT, Dudina A, D'Agostino R, Graham IM. Cardiovascular risk-estimation systems in primaryprevention: do they differ? Do they make a difference?Can we see the future?Circulation. 2010; 122(3):300–10.doi: 10.1161/CIRCULATIONAHA.109.852756 PMID: 20644026.

23. AssmannG, Cullen P, Schulte H. Simple scoring scheme for calculating the risk of acute coronaryevents based on the 10-year follow-up of the prospective cardiovascular Munster (PROCAM) study.Circulation. 2002; 105(3):310–5. PMID: 11804985.

24. Diener A, Celemin-Heinrich S, Wegscheider K, Kolpatzik K, Tomaschko K, Altiner A, et al. In-vivo-vali-dation of a cardiovascular risk prediction tool: the arriba-prostudy. BMC Fam Pract. 2013; 14:13. doi:10.1186/1471-2296-14-13 PMID: 23339773; PubMedCentral PMCID: PMC3583804.

25. Gohlke H,WinterM, Karoff M, Held K. CARRISMA: a new tool to improve risk stratificationand guid-ance of patients in cardiovascular riskmanagement in primaryprevention. Eur J Cardiovasc PrevRehabil. 2007; 14(1):141–8.doi: 10.1097/01.hjr.0000244581.30421.69 PMID: 17301640.

26. Muller-Riemenschneider F, Holmberg C, Rieckmann N, KliemsH, Rufer V, Muller-Nordhorn J, et al.Barriers to routine risk-scoreuse for healthy primarycare patients: survey and qualitative study. ArchInternMed. 2010; 170(8):719–24. doi: 10.1001/archinternmed.2010.66PMID: 20421559.

27. Collaborators GBDRF, Forouzanfar MH, Alexander L, AndersonHR, Bachman VF, Biryukov S, et al.Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occu-pational, andmetabolic risks or clusters of risks in 188 countries, 1990–2013: a systematic analysis forthe Global Burden of Disease Study 2013. Lancet. 2015. doi: 10.1016/S0140-6736(15)00128-2 PMID:26364544.

28. Fine JP, Gray RJ. A Proportional HazardsModel for the Subdistribution of a Competing Risk. Journal ofthe American Statistical Association. 1999; 94(446):496–509.

29. Cooney MT, Selmer R, LindmanA, Tverdal A, Menotti A, ThomsenT, et al. Cardiovascular risk estima-tion in older persons: SCOREO.P. Eur J Prev Cardiol. 2015. doi: 10.1177/2047487315588390 PMID:26040999.

Risk Chartsof Cardiovascular Disease in Germany

PLOSONE | DOI:10.1371/journal.pone.0162188 September 9, 2016 11 / 11